Re: regression (crash) in sysmon/acpiacad

[Jukka Ruohonen <jruohonen%iki.fi@localhost>]

On Sun, Feb 07, 2010 at 08:30:27AM +0100, Joerg Sonnenberger wrote:
> On Sun, Feb 07, 2010 at 09:04:54AM +0200, Jukka Ruohonen wrote:
> >     * The following sensors should be removed: technology,
> >       low capacity, and warning capacity. These are not really
> >       something that should be "sensed".
> 
> Technology ok. I'm not too sure about low and warning, given that they
> normally can't be modified.

The idea here would be to use the sme_get_limits() and possibly
sme_set_limits(). This is exactly the rationale behind those callbacks.

This would also result a nicer output in envstat(8).

> 
> >     * The design capacity should be the maximum of the last known full
> >       charge capacity, which is the maximum of the present capacity. 
> >       This is useful for checking the overall "health" of deteriorating
> >       (lithium-ion) batteries.
> 
> I disagree. Both batteries for my laptop had initially a higher capacity
> than designed for -- e.g. last full and design cap don't necessarily
> agree with each other.

I noticed the same thing with voltages. Yet, what is wrong with envstat(8)
or some other tool reporting "last full charge capacity is 123 % of the
design capacity"?

> >     * Sensors that have a maximum should report also percentages in
> >       relation to these maximums. From the usability point of view, this
> >       is probably almost always the right choice.
> 
> That should be a task for userland, not the kernel.

It already is; in acpibat(4) this just implies setting the ENVSYS_FPERCENT
flag, nothing more.

- Jukka.